Sie sind auf Seite 1von 15

Professor Ahmed DRIOUCHI, PhD

Institute of Economic Analysis &


Prospective Studies (IEAPS),
Al Akhawayn University, Ifrane, Morocco
Senior Lecturer Cristina BOBOC, PhD
The Bucharest Academy of Economic Studies, Romania

DO BETTER HEALTH AND EDUCATION ENHANCE ECONOMIC


DEVELOPMENT?

Abstract. The major objective of this study is to show the importance of


interdependencies between health, education and poverty in the context of
South Mediterranean Countries (SMC) compared to the economies of the
European Union (EU). Empirical assessments of interdependencies are pursued
after reviewing the previous theories and applications. The results attained have
revealed the existence of important links between different sources of wealth
which are health, knowledge, monetary assets and social capital. Furthermore,
the importance of non monetary assets appeared to be at least as important as
the traditional economic and financial wealth. Given the extent and magnitude
of education, health and socio-economic deficits in the SMC, the identification
of the interdependencies generates new avenues for more integrated economic
and social policies.
Keywords: Interdependencies, Health, Education, Poverty, South
Mediterranean Countries.

JEL Classification D31-I1-I2-I3

1. INTRODUCTION

This study aims at analyzing the interdependencies existing between


health, education and poverty as preliminary steps towards an analytical focus on
human development policies. Previous research already demonstrated the role of
knowledge and education in driving economic performance (Driouchi, A. et al,
2006). Yet, how do other variables, such as health, that are directly related to
human development interact and affect economic and social development?
If empirical research has been accumulated globally and on different regions of the
world, only limited evidence on the area of interdependencies exists for the South
Mediterranean economies. The Arab human development reports (UNDP, 2002, 2003,
2004 and 2005) have shown large gaps in different aspects of human development in
Ahmed Driouchi, Cristina Boboc
the region. Health and knowledge have appeared to be among the major deficits in
the region and occupy key positions in limiting access to better lives.
Insights into these inter-related issues have been building with the development of
both theoretical and empirical studies, focusing on interdependencies with the
development of new challenges facing policy-makers when accounting for the
magnitude and level of sector inter-relations.
As developed in different empirical reports and studies, human development policies
do have important interdependencies that should be better identified. These
identifications help in refining the knowledge and the direction of policy actions
that could enhance the level of attainment of the Millennium Development Goals
(MDG’s), as well as the enhancement of the living conditions of the populations.
This research has two inter-related objectives. The first one is to show how health,
education, and income poverty are inter-related in the context of the South
Mediterranean Countries compared to EU countries. The second objective is to
relate health and education variables of the region and to identify how they can
contribute to enhance economic development.
To achieve the above objectives, the most recent theoretical economic models are
reviewed and used to support the empirical investigations that are pursued in the
context of this study. Different databases from international organizations have
provided the variables and indices included in the empirical investigations.
Descriptive and regression analysis are used to assess and test for the existence of
interdependencies between health, education and earnings.

2. LITERATURE REVIEW ON INTERDEPENDENCIES OF


HEALTH, EDUCATION AND WEALTH
In an earlier contribution, Driouchi A. et al. (2006) presented the relationship
between knowledge and economic performance using panel data on both developed
and developing economies. Among the results attained, human capital and mainly
education demonstrated promising roles in the determination of economic
performance. However, other human capital related variables, such as health
components, were not included given the methodological framework used in the study.
Insights into the likely relationships of health, education and economic performance
have been explicitly discussed in a series of social science and public health
publications and reports. Different approaches have been used to tackle the extent
and magnitude of these relationships. It is obvious that these linkages are well
known, but their magnitude and extent of their usefulness for economic and social
policies are not often emphasized. This review deals with the importance of
interdependencies and the major findings accumulated in the socioeconomic literature.
For Bernard et al. (2006), three basic sources of wealth (health, education, and
economic resources) can be seen as interdependent causes and consequences of one
another over the lifetime of an individual. The social capital could be considered as
an important factor because individuals can accumulate other useful resources for
themselves, their families and their communities through social networks.
Do Better Health and Education Enhance Economic Development?
Adams et al. (2003) identified causal relationships between “socioeconomic status”
and health conditions, with the empirical evidence established between
socioeconomic factors and health but no clear causality found.
Lee and Kim (2007) conducted a longitudinal analysis to detect the long-term
effect of health shocks on wealth and compared that with its short-term effect on
the elderly, which is consistent with the previous studies including that of Adams
et al. (2003). New health events appeared to have negative impacts on wealth but
disappeared over time (Lee and Kim, 2007). The results also verified that severe
health conditions (existing and new) significantly influenced wealth depletion
mainly when shocks happen later in life. Furthermore, the study confirmed that
health capital (existing severe chronic conditions) has a persistent negative impact
on wealth changes over time. These results are subject to variation with the level of
education, the family status, and other factors. Longer term effects were also
investigated using panel data by P. Adams et al (2002) and by Cutler et al. (2007).
However, limited evidence has been proven; the same results were found through
the study of Cutler, Miller and Norton (2007). Mayer-Foulkes (2004) addressed the
long-term impact of health by including the intergenerational and life-long
dimensions. The relationships between health and each of the components that
define the wealth of an individual, a group, and a country are examined below
using a variety of publications.
Cutler David M, Adriana Lleras-Muney and Tom Vogl (2008) presented the
existence of a clear link between socioeconomic status and health, which are
identified by a number of studies covering both the United States and European
Countries. For example, mortality risk rises when individuals do not reach upper
secondary education in the United States and in some European Countries. Each of
the measures of socioeconomic status influences health through different
mechanisms. This has important implications for the choice of public policies
aiming at improving the health of individuals under specific conditions.
Sommestad (2001) emphasized the role of investment in human capital, with health
being the major engine of economic growth. Based on the empirical evidence, “a
5-year gain in life expectancy resulted in 0.3 to 0.5 per cent economic growth”.
Poverty and poor health penalize economic growth (Bloom and Canning, 1999). In
The Health and Wealth of Nations, Bloom & Canning (2000) acknowledged the
relationship between health and income, which indicates that higher income leads
to a longer life expectancy. A healthier population works more efficiently with
higher chances of improving its skills, generates and attracts more investments, and
benefits from a higher resource allocation. This causality between health and
income lead to health improvement and, then to further income increase. Various
examples of this “Virtual Spiral” are from East Asia and Ireland (Bloom & Canning,
2000). In another publication, the same authors (2004) stressed that investing in
health leads to higher economic and social performance under sound
macroeconomic policies and governance. The same study illustrated that initial
beneficiaries of health improvements are often the most vulnerable groups (children),
Ahmed Driouchi, Cristina Boboc
with healthier children having better school attendance and improved performance.
This shows how health and income have been identified to be highly related.
Hurd and Kapteyn (2005) also analyzed the relationship between health and income.
They found that, in some countries, a large variation in wealth is associated with a
large variation in health. Further techniques and analysis have shown the
importance of the link between health and expenditures. Xu et al. (2003) demonstrated
the existence of an overall positive relationship between the proportion of
households with catastrophic health expenditures and the share of out-of-pocket
payments in total health expenditures. Gerdtham and Thgren’s (2002) found that
health expenditure and GDP are co-integrated around linear trends.
Knapp (2007) dealt with the links between nutrition, labor productivity and a
health variable, height. The net role of nutrition on labor productivity was shown to
be highly significant. Muysken et al. (1999) showed that when physical capital,
relative to health, is scarce, optimal expenditures for health and consumption are
lower. Also, consumption is no longer likely to be negatively related to population
growth because it enhances the percentage of health workers.
Wichmann (1995) investigated the nutrition-productivity relationship at low levels
of income and showed the existence of a strong relationship between the state of
nutrition or health and labor productivity. The dynamic version demonstrated that
better nutrition increases the productivity of the learning by doing process. The
empirical evidence gathered confirmed the above findings, but revealed that
children with good nutrition and health start school at an earlier age, progress
further, and repeat fewer grades (Yamauchi, 2006). However, the analysis of a long
panel of data (11 years) suggests that good health may discourage further
investments in schooling at the stage of transition from primary to secondary
school and that better health status may reinforce incentives to go to work.
A further series of publications produced major inputs into how interdependencies
could be used to support more realistic policies. Anderson et al. (2004) illustrated how
to account for evidence about early childhood socioeconomic conditions, which have
long-term health consequences on health disparities over the life course, in relation to
early investments in education. Aging populations in the developed world are currently
posing a serious threat to the cost of pensions and health care. Anderson et al. (2004)
explained why countries cannot expect to grow themselves out of the problem
using measures such as increasing immigration, raising the retirement age, and
encouraging households to have more children.
According to Farrell et al. (2005), the only effective measures are for households and
governments to increase their savings rates and to allocate capital more efficiently in
order to earn higher returns on the assets they have.
Grimm and Harttgen (2007) looked at the role of the dependency ratio, in relation to
population growth, in relation to health. They also found that a low life expectancy
substantially reduces welfare, despite the related economic feedback effects.
Do Better Health and Education Enhance Economic Development?
3. EMPIRICAL INVESTIGATIONS
There are compelling reasons to believe that education and health positively
contributes to economic growth. Yet, empirical studies show that this relationship
is not always significant. Barro (1990) demonstrated that, for a given level of
wealth, the economic growth rate was positively related to the initial level of
human capital of a country; whereas, for a given level of human capital, the growth
rate was negatively related to the initial level of GDP per capita. Azariadis and
Drazen (1990) assumed that economic growth is not a linear process. Their results
showed that literacy is correlated with the variation of growth in the least advanced
countries, but it does not seem to be related to the growth of the most developed
countries. Furthermore, they illustrated that the coefficient of human capital in the
growth equation is about five times higher in the developing countries than in the
developed countries. Therefore, one of the main conclusions of the analysis of the
human capital-growth relationship is the absence of homogeneity across countries. If
economic, social and cultural characteristics of each country modify the relationship
between human capital and growth, then the estimation of this relationship must be
regional and take into account the initial level of human capital. In order to include
initial wealth of a country as factor in the econometric analysis, a dummy variable was
created, which indicates if a country is poor (dummy=0) or not (dummy=1).
This analysis is based on datasets made of World Bank and United Nations. The
first part uses descriptive statistic methods and simple regressions to make a
comparative analysis between SMC and EU countries about the levels of
education, health and economic wealth. The second part uses factorial methods to
detect the interdependencies between education, health and economic development.
Using these variables, three wealth patterns are observed and are analyzed for SMC
and EU countries. The last part is an econometric analysis of the relationship between
education, health and economic development using models with dummy variables.

3.1 Descriptive statistics


Per capita economic development in the SMC region in the past 20 years has been
relatively low. This is partly because of high population growth rates, and partly
because many countries still depend on oil exports, and oil prices remained relatively
low during this period. Under this situation, human capital (health and education) and
economic development are not expected to show meaningful relationships.
The South Mediterranean region is a medium wealthy region. During the period
1995-2007, the annual GDP growth was 3.87% for the region, with high variation
between countries. Djibouti has the lowest annual GDP growth of 1.35% on
average per year, and United Arab Emirates has the highest annual GDP growth of
6.75% on average per year.
Many scientists consider that literate people have better health and a higher socio-
economic status. The general measure used to describe the educational level of a
country is the literacy rate, which is calculated as the percentage of the population
15 years and older who are literate (can read and write). The variation of literacy
Ahmed Driouchi, Cristina Boboc
rate across countries is very high, from 52% for underdeveloped countries to 94%
for developed ones.
The SMC region has invested heavily in education over the past few decades. As a
consequence, the mean of the literacy rate increased from 68% in 1995 to 79% in
2004. However, disparities between countries are very high.
Life expectancy at birth is one of the most important indicators of health. For much
of human history, life expectancy was between 20 to 35 years (Preston, 1995). In
2004, it was 72 years for the region under study, compared to 66 years worldwide.
Life expectancy at birth has increased with the level of development, from 47 years
for the underdeveloped countries to 77 years for the developed countries.
Recent studies argue that correlations of education with health and socio-economic
status may have more to do with the effects of schooling rather than literacy in
general. School life expectancy shows the overall level of development of an
educational system in terms of the number of years of education that a child can
expect to achieve. Special attention is required in the interpretation, as long as
relatively higher school life expectancy indicates greater probability for children to
spend more years in education but also higher overall retention within the
education system. The overall mean of school life expectancy is 12 years, much
higher in developed countries than in developing or underdeveloped countries (15
years in developed countries versus 7 years in underdeveloped countries). In the
region, school life expectancy increased from 11 years, in 1995, to 12 years in 2004.
Between countries, there are large disparities. For example, school life expectancy in
Djibouti was 4 years in 2004, compared to Israel where it was 15 years.

3.2 How education, health and economic development are interrelated?


In order to have a suggestive representation of the interrelations between education,
health and economic development, Principal Component Analysis (PCA) is applied.
The variables used in this analysis are: GDP per capita PPP (current international
dollars), life expectancy at birth (years), adult literacy rate, infant mortality rate (per
1,000 live births), and school life expectancy (years) - primary to tertiary, for SMC and
EU countries, for three years 1995, 2000 and 2004. As long as adult literacy rate is
available only for the year 2004, it has been considered constant for all these years.
Most of the information is preserved (92%) by the projection of the variables on the
plan determined by the first two principal components.
On the first axis, the best represented variables are on one side: school life
expectancy (years) - primary to tertiary and adult literacy rate, and on the other
side: infant mortality rate (per 1,000 live births). The first axis summarizes the
educational and health components of wealth. School life expectancy (years) -
primary to tertiary and adult literacy rate are strongly positively correlated, and
negatively correlated with infant mortality rate. The countries on the positive side
of the first axis have values of educational indicators higher than the average, and
the values of infant mortality rate lower than the average. The farther the country is
from the center of the axis on the positive side, the wealthier in education and
Do Better Health and Education Enhance Economic Development?
healthier the country is. The farther the country is from the center of the axis on the
negative side, the poorer in education and the lower in health the country is. The
SMC countries with the highest levels of education and lowest levels of infant
mortality rates are Israel and Jordan, and the countries with the lowest levels of
education and highest levels of infant mortality rates are Yemen and Morocco.
Countries like Morocco, Yemen, Jordan and Tunisia became richer in education
and health during the last ten years, while countries like Israel, Saudi Arabia, Iran,
Oman and Kuwait became poorer in education and health during the last ten years.

Component Plot in Rotated Space

1.0 GDP
LEB

0.5
SLE
LR
Component 2

0.0

-0.5 MR

-1.0

-1.0 -0.5 0.0 0.5 1.0


Component 1

Figure 1 – Variable representation on factorial plan

On the second axis, the best represented variables are on one side: GDP per capita
PPP (current international $) and life expectancy at birth (years), and on the other
side infant mortality rate (per 1,000 live births). The second axis summarizes the
economic and health components of wealth. School GDP per capita and life
expectancy at birth are strongly positively correlated, and negatively correlated
with infant mortality rate. The countries on the positive side of the second axis
have values of GDP per capita and life expectancy at birth higher than the average
and the values of infant mortality rate lower than the average. The farther the
country is from the center of the axis on the positive side, the richer and healthier
the country is. The farther the country is from the center of the axis on the negative
side, the poorer the country is. The richest and healthiest SMC countries are Israel
and United Arab Emirates, and the poorest country is Yemen. During the last ten
years, all of the countries became richer and healthier.
Ahmed Driouchi, Cristina Boboc

2.00000

UNITED ARAB EMIRATES, 2000 ISRAEL, 2004 Cluster 3


KUWAIT, 2004
ISRAEL, 2000
REGR factor score 2 for analysis 1

UNITED ARAB EMIRATES, 1995BAHRAIN, 2004


1.00000
OMAN, 2004 ISRAEL, 1995

KUWAIT, 1995
SAUDI ARABIA, 2004

OMAN, 2000
SAUDI ARABIA, 2000

MOROCCO, 2004
SAUDI ARABIA, 1995
BAHRAIN, 1995
Cluster 1
0.00000 TUNISIA, 2004
MOROCCO, 2000 TUNISIA, 2000
MOROCCO, 1995 ALGERIA, 2004
IRAN, 2004

TUNISIA, 1995
IRAN, 2000
JORDAN, 2004
IRAN, 1995
-1.00000 JORDAN, 2000

YEMEN, 2004

YEMEN, 2000
Cluster 2
YEMEN, 1995
-2.00000

-3.00000 -2.00000 -1.00000 0.00000 1.00000 2.00000


REGR factor score 1 for analysis 1

Figure 2 – SMC countries representation on factorial plan and wealth clusters

2.00000 ITALY, 2004


Cluster 3
SPAIN, 2004
ITALY, 2000
CYPRUS, 2004
CYPRUS, 2000 GREECE, 2004
REGR factor score 2 for analysis 1

MALTA, 2000 SPAIN, 2000


1.00000 MALTA, 2004
ITALY, 1995
PORTUGAL, 2004
GREECE, 2000
CYPRUS, 1995 SLOVENIA, 2004
PORTUGAL, 2000 SPAIN, 1995
SLOVENIA, 2000 Cluster 1
GREECE, 1995
PORTUGAL, 1995

0.00000
SLOVAK REPUBLIC, 2004
SLOVENIA, 1995
SLOVAK REPUBLIC, 2000

ESTONIA, 2004
BULGARIA, 2004 LITHUANIA, 2004
SLOVAK REPUBLIC, 1995 LATVIA, 2004
ROMANIA, 2004 LITHUANIA, 2000
ROMANIA, 2000 ESTONIA, 2000
-1.00000
BULGARIA, 2000 LATVIA, 2000
BULGARIA, 1995
ROMANIA, 1995 LITHUANIA, 1995

ESTONIA, 1995
Cluster 2 LATVIA, 1995

-2.00000

-3.00000 -2.00000 -1.00000 0.00000 1.00000 2.00000


REGR factor score 1 for analysis 1

Figure 3 – EU countries representation on factorial plan and wealth clusters


Do Better Health and Education Enhance Economic Development?
Taking into consideration all of the variables, the wealthiest countries in health, in
education and economically will be represented in the first dial, far from the axis
center, and the poorest will be represented in the third dial, far from the axis center.
Thus, the richest SMC country from all points of view is Israel, and the poorest
SMC country from all points of view is Yemen.
The main differences between SMC countries and EU countries are in education.
Almost all SMC countries are on the negative side of the first axis, and almost all
EU countries are on positive side of the first axis, which is determined by
educational wealth and infant mortality rate. The general time trend is the
translation of countries to higher values of education, health and GDP. For many of
SMC countries, the improvement in education is more visible; for most of EU
countries, the improvement in GDP per capita and health is more important. The
homogeneity on the first axis is much higher for EU countries than for SMC
countries; so, EU countries are much more homogenous in education and infant
mortality rate compared to SMC countries.

3.3 Wealth patterns by education, health and economic development


This section applies data clustering methods in order to determine the wealth
patterns by education, health and economic development for SMC countries
compared to wealth patterns for EU countries. Data clustering means the
classification of objects into different groups (clusters), so that the data in each
cluster share some common attributes, often proximity according to some defined
distance measure. In this case, the distance measure is the Euclidean distance, and
the methods used are Ward method and k-means clustering method. Three groups
are identified having the following characteristics:
• Cluster 1 includes countries with medium wealth: Bahrain, 1995; Kuwait,
1995; Saudi Arabia, 1995, 2000, 2004; and Oman, 2000, 2004. The school life
expectancy (primary to tertiary) mean is 13.6 years, and the adult literacy rate
mean is 92%. Hence, this group includes countries with average levels of
education. The life expectancy at birth mean is 74.6 years, and the infant
mortality rate mean is of 9.8 dead infants per 1,000 live births. Therefore, this
cluster includes countries with an average level of health. The GDP per capita,
PPP mean is 14,663 current international $.
• Cluster 2 includes poor countries from all points of view: Algeria, 2004; Iran,
1995, 2000, 2004; Jordan, 2000, 2004; Morocco, 1995, 2000, 2004; Tunisia,
1995, 2000, 2004; and Yemen, 1995, 2000, 2004. The school life expectancy
(primary to tertiary) mean is 12.13 years, and the adult literacy rate mean is 82%.
Hence, the countries from this group have low levels of education. The life
expectancy at birth mean is 69.2 years, and the infant mortality rate mean is
30.2 dead infants per 1,000 live births. Therefore, this cluster includes
countries with low level of health. The GDP per capita, PPP mean is 5,649
current international $.
Ahmed Driouchi, Cristina Boboc
• Cluster 3 includes rich countries: Bahrain, 2004; Israel, 1995, 2000, 2004;
Kuwait, 2004; and United Arab Emirates, 1995, 2000. The school life
expectancy (primary to tertiary) mean is 14.46 years, and the adult literacy rate
mean is 95%. Hence, the countries from this group have average levels of
education. The life expectancy at birth mean is 78 years, and the infant
mortality rate mean is 5.9 dead infants per 1,000 live births. Therefore, this
cluster contains countries with an average level of health. The GDP per capita,
PPP mean is 22,171 in current international dollars.

During the analyzed period, some translations could be observed from one cluster
to another. Kuwait and Bahrain passed from cluster 1 to cluster 3, from medium
wealthy countries to rich countries. Many of the new entrants in EU passed from
cluster 2, countries with low level of wealth indicators, to cluster 1, countries with
medium level of wealth indicators. The classification of countries and the changes
in time confirm the interdependencies between the wealth components. The
increase in GDP is correlated with the improvement in health and in education.
SMC countries have lower education level and lower level of health compared to
EU countries for the same level of GDP per capita.

3.4 What are the main determinants of wealth patterns?


In order to answer this question, a discriminant function analysis is used. The
results obtained are significant, with a probability of 95%. Discriminant analysis
determines some optimal combinations of variables (called discriminant functions)
so that the first function provides the most overall discrimination between groups,
the second provides second most, and so on. Moreover, the functions will be
independent or orthogonal, that is, their contribution to the discrimination between
groups will not overlap. In this case, the first function discriminates between the
first cluster and the others, so between the medium wealth countries and the rest of
the countries. The variables which contribute the most to this discrimination are
GDP per capita and life expectancy at birth. This function is significant with a
probability of 99%. The second function discriminates between the second and the
third cluster, so between the rich and poor countries. The variables which contribute
the most to this discrimination are school life expectancy, adult literacy rate and
infant mortality rate. This function is significant with a probability of 90%.
Therefore, GDP per capita and life expectancy at birth determine the pattern of
medium wealth countries. To be identified as a rich or as a poor country, the other
variables, such as school life expectancy, adult literacy rate and infant mortality
rate, are important. Thus, the economic and health components of wealth are used
for the first time in identifying the wealth patterns of a country.
By using discriminant analysis, three classification functions are determined:
Cluster 1: f1 = -0.016·GDP + 41.518·LExp + 7.24·LR + 10.433·MRI + 2.375·SLE - 1832.509
Cluster 2: f2 = -0.019·GDP + 42.559·LExp + 7.329·LR + 10.574·MRI + 3.54·SLE – 1903.387
Cluster 3: f3 = -0.014·GDP + 41.157·LExp + 7.321·LR + 10.592·MRI + 1.564·SLE – 1849.568
Do Better Health and Education Enhance Economic Development?
One country is classified into the group with the highest value of the classification
function. By using these classification functions, 100% of the cases used in the
analysis are correctly classified.
Some simulations for the SMC countries are made in order to check the possibility
of passing from one group of countries to another (from a medium wealthy country
to a rich country of from a poor to a medium wealthy country), improving only one
component of wealth. The simulation sought to find out if changing only one
component of wealth could make a significant impact, or if it is necessary to
improve more or less all components of wealth. For the four SMC countries
(Kuwait, Jordan, United Arab Emirates, Morocco), the values for the classification
functions were calculated, varying one by one the indicators from the minimum to
the maximum observed value in the database and maintaining constant all the other
indicators at the 2004 observed level. Improvements in GDP per capita involve the
revision of country classification. The speed of change depends on the level of all
the other wealth components. The higher the levels of health and education are, the
faster the adjustment to a better group of countries is.

The main conclusion of this analysis is that all the wealth indicators are
interdependent. Modifying or improving only one indicator is not sufficient for a
country to become wealthier. However, a better health of population determines a
better human capital, so more productivity, and in conclusion, a better economic
result at country level. More educated people will determine an improvement in
human capital, which leads to a greater level of GDP per capita. Furthermore, a
richer country, from an economic point of view, will invest more in education and
in the health of people, but the response is not as quick as for the other
implications. The next two sections give some measurable results of these two
statements.

3.5 Econometric models


In order to determine the impact of human capital on economic development, an
econometric analysis with dummy variables is performed. The variables included
in the analysis are: GDP growth (%), life expectancy at birth (years) and school
life expectancy (years) - primary to tertiary, for SMC and EU countries, for years
1999, 2003 and 2005 or the last year available. In order to include initial wealth of
a country as a factor in the econometric analysis, a dummy variable is created,
which indicates if a country is poor (dummy=0) or not (dummy=1), using the
previous results obtained by factorial analysis for the years 1995, 2000 and 2004.

The effect of Health on Economic Development


Regarding the relationship between health and GDP growth, there is a strong
relationship between them. All of the models are very good with R2 greater than 90%
for both SMC countries and EU countries. The parameter coefficients are
significant with a probability of at least 90%.
Ahmed Driouchi, Cristina Boboc
The parameter estimates of the regression models, which describe the relationship
between health and GDP growth taking into account the initial level of wealth of a
country, are provided in table 1. The main conclusion provided by these models is
that the effect of improvements in health on GDP growth is more important for rich
or medium wealthy SMC countries than for poor SMC countries.
An increase of 1 year in life expectancy at birth implies an increase of GDP growth
2 times higher for poor SMC countries than for the other SMC countries. The
equation for poor countries goes through the origin and for the other countries
there is a translation of the regression line with 4.441. Concerning poor EU
countries, the impact of the increase in life expectancy at birth on GDP growth is
much higher than for the poor SMC countries. In case of the not poor countries,
there is an increase in GDP growth with life expectancy at birth for SMC countries,
but not for EU countries, where there is an inverse dependency (Fig. 4).

Table 1 - Econometric models describing the effect of health on economic


development
Variables** N R2 F Model
SMC countries
Dependent variable: GDP 29 0.998 3812 Ln(GDPg) = 0.067·LEB +
Independent variables: LEB (76.81)
Dummy variable + 4.441·Dummy - 0.065·LEB·Dummy
(2.54) (-2.81)
EU countries
Dependent variable: GDP 33 0.99 225305 Ln(GDPg) = 0.065·LEB +
Independent variables: LEB (474.5)
Dummy variable + 5.139·Dummy -0.071·LEB·Dummy
(24.304) (-26.023)
*Under each regression coefficient, the corresponding t-value is written
** Dummy=0 if poor country and Dummy=1 if not poor country; GDPg = GDP growth+100%; LEB = Life
expectancy at birth, total (years)

SMC Countries EU countries


900 700
800 600
700
500
GDP growth

GDP growth

600
500 400
400 300
300
200
200
100 100
0 0
1
6
11
16
21
26
31
36
41
46
51
56
61
66
71
76
81
86
91
96

1
6
11
16
21
26
31
36
41
46
51
56
61
66
71
76
81
86
91
96

LEB LEB
poor countries medium and rich countries poor countries medium and rich countries

Figure 4 – Regression curves (GDP growth versus life expectancy at birth)


Do Better Health and Education Enhance Economic Development?
The effect of Education on Economic Development
Concerning the relationship between education and GDP growth, there is a strong
relationship between them. The models are good with R2 greater than 98%. The
parameter coefficients are significant with a probability of 99%.
The parameter estimates of the regression models, which describe the relationship
between education and GDP growth taking into account the initial level of wealth
of a country, are provided in table 2. The impact of the increase in school life
expectancy on GDP growth is much higher for poor SMC countries than for the poor
EU countries. In case of the not poor countries, GDP growth decreases faster with
school life expectancy increases for SMC countries than for EU countries (Fig. 5).

Table 2 - Econometric models describing the effect of education on economic


development
Variables** N R2 F Model
SMC countries
Dependent variable: GDP 25 0.982 404.3 Ln(GDPg) = 0.396·LEB +
Independent variables: LEB (25.866)
Dummy variable + 4.711·Dummy - 0.401·LEB·Dummy
(3.038) (-3.41)
EU countries
Dependent variable: GDP 33 0.998 4615.6 Ln(GDPg) = 0.337·SLE +
Independent variables: SLE (67.77)
Dummy variable + 4.663·Dummy -0.338·SLE·Dummy
(7.197) (-7.897)
*Under each regression coefficient, the corresponding t-value is written
** Dummy=0 if poor country and Dummy=1 if not poor country; GDPg = GDP growth+100%; SLE = School life
expectancy (years). Primary to tertiary. Total

SMC EU
1400 500
450
1200
400
1000 350
GDP growth

GDP growth

800 300
250
600
200
400 150
200 100
50
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

SLE SLE
poor countries medium and rich countries poor countries medium and rich countries

Figure 5 – Regression curves (GDP growth versus school life expectancy)

4. CONCLUSION
The regression analysis of different clusters, concerning the potential relationships
between education, health and income measurements, helps to come up with the
following conclusions.
Ahmed Driouchi, Cristina Boboc
First, the use of aggregate data, such as the “world cluster”, to illustrate
relationships, without taking care of the different economic, social, regional, and
development specificities, leads to contradictory findings. The example of the
regression analysis of children out of primary school illustrates the ineffectiveness
of aggregate data, because the results on world data contradict completely the
results of the five sub-clusters. Data from sub-clusters, in contrast, improve the
efficiency and help detect the level of significance existing between the GDP per
capita and both health and education.
Second, the GDP per capita impact on improving education is not very significant.
In fact, the overall conclusion that could be drawn from the regression analysis of
the education and income indicators is that the relationship significance of GDP
per capita is not considerable. In other words, the improvement of education within
different clusters does not necessarily require an economic involvement or an
increase in the income of the population. The single argument in favor of the
importance of the GDP per capita for ameliorating the level of education concerns
the literacy rates in middle income countries (mainly SMC, major oil net exporters,
and high and medium human development). In general, other arguments besides
income should be studied in order to enhance the contribution of education.
Third, in contrast to education, the interpretations of regression tables for health
and income indicators show that GDP per capita is the key for most clusters. The
enhancement of health in different clusters can only occur if it is accompanied by
an improvement of the income level of the population. Alternatively, low human
development countries need to focus on other measures besides GDP per capita to
determine a significant relationship with health issues. The main point of the low
human development countries’ health indicators is that the relationship
significance varies from year to year. This “instability” can be explained by the rise
of new factors that alter health, more than the GDP per capita would do, within a
particular period.
Fourth, the economic development depends significantly on improvements in
education and health. The impact of human development on GDP growth depends
on the initial wealth of the country and on region (SMC or EU).

REFERENCES

[1] Adams, et al. (2003), Healthy, Wealthy and Wise? Tests for Direct Causal Paths
between Health and Socioeconomic Status, Journal of Econometrics, 112: 3-56;
[2] Barro, R.J. (1991), Economic Growth in a Cross-Section of Countries,
Quarterly Journal of Economics, 106: 407-443.
[3] Bloom, D.E., Canning D., Sevillahttp J. (2001), The Effect of Health on
Economic Growth: Theory and Evidence. Working Paper 8587, National
Bureau of Economic Research http://www.nber.org/papers/w8587.pdf
[4] Cutler, D.M., Lleras-Muney A., Vogl T. (2008), Socio Economic Status and
Health: Dimensions and Mechanisms, NBER Working Paper 14333;
Do Better Health and Education Enhance Economic Development?
[5] Driouchi, A. et al. (2006), An Econometric Analysis of the Role of
Knowledge in Economic Performance, Journal of Technology Transfer, 31,
241–255, Springer Science Business Media, Inc. Manufactured in The
Netherlands;
[6] Farrell, D., Ghai S., Shavers T. (2005), The Coming Demographic Deficit:
How Aging Population Will Reduce Global Wealth. March ;
[7] Grimm, M., Harttgen K. (2007) , Longer Life, Higher Welfare?. Oxford
University Press 1 of 19, doi:10.1093/oep/gpm025;
[8] Hurd, M., Kapteyn A. (2005), Health, Wealth and the Role of Institutions,
in Multidisciplinary Economics: 307-332, P.fr Gisjel and H. Schenk editors,
Springer, The Netherlands;
[9] Lee, J. and Kim H. (2007), A Longitudinal Analysis of The Impact of
Health Shocks on The Wealth of Elders . Journal of Population Economics;
[10] Mayer-Foulkes, D. (2004) , The Intergenerational Impact of Health on
Economic Growth. CIDE. Written for the Global Forum for Health Research,
Forum 8, Mexico City, 16 -20 November;
[11] Muysken, J., Hakan Yetkiner I. and Ziesemer T. (1999), Health, Labour
Productivity and Growth. CCSO Working Papers 200015, University of
Groningen, CCSO Centre for Economic Research, revised;
[12] Scarlat E., Maracine V. (2008) , Agent-based Modelling of Forgetting
and Depreciating Knowledge in a Healthcare Knowledge Ecosystem,
Journal of Economic Computation and Economic Cybernetics Studies and
Research, No. 3-4, ASE Publishing House, Bucharest;
[13] Sen, A. (1998) , Mortality as an Indicator of Economic Success and
Failure. Economic Journal, Royal Economic Society, vol. 108(446),
pages 1-25;
[14] Sommestad, L. (2001), Health and Wealth: The Contribution of Welfare
State Policies to Economic Growth. Institute for Futures Studies. Speech
prepared for the Expert Conference" Best Practices in Progressive
Governance”.
[15] Wichmann, T. (1995) , Food Consumption and Growth in a Two Sector
Economy. Technical University Berlin;
[16] The World Bank (2008) , The Road not Traveled. Education Reform in
Middle East and Africa, MENA Development Report.